Update app.py

app.py

@@ -1,6 +1,7 @@
 import os
 import gradio as gr
 import requests
+import json
 
 # Get the Hugging Face API key from Spaces secrets.
 HF_API_KEY = os.getenv("HF_API_KEY")
@@ -13,39 +14,39 @@ MODEL_ENDPOINTS = {
 }
 
 def query_model(prompt, model_endpoint):
-    """
-    Query a model via the Hugging Face Inference API using a requests.post call.
-    This payload now uses the 'inputs' field as required.
-    """
     headers = {
         "Authorization": f"Bearer {HF_API_KEY}",
-        "Content-Type": "application/json"
+        "Content-Type": "application/json",
+        "Accept": "application/json"
     }
-    # Use 'inputs' instead of 'messages' and pass parameters inside the 'parameters' field.
+    # Format prompt for instruct-style generation
+    formatted_prompt = f"User: {prompt}\nAssistant:"
     data = {
-        "inputs": prompt,
+        "inputs": formatted_prompt,
         "parameters": {
             "max_new_tokens": 512,
             "temperature": 0.7,
         }
     }
-    
     response = requests.post(model_endpoint, headers=headers, json=data)
     
+    # For debugging purposes, you can uncomment the following line:
+    # print("Raw response:", response.text)
+    
     try:
         result = response.json()
     except Exception:
         return f"Error: Unable to parse JSON. Response: {response.text}"
     
-    # If there's an error message in the result, return it.
+    # If the API returns an error message, surface it.
     if isinstance(result, dict) and "error" in result:
         return f"Error: {result['error']}"
     
     try:
-        # Expected result format is a list of outputs with a "generated_text" field.
-        return result[0]["generated_text"]
+        # Expecting a list of outputs with a "generated_text" field.
+        return result[0].get("generated_text", "No generated_text found in response")
     except Exception:
-        return f"Error: Unexpected response format: {result}"
+        return f"Error: Unexpected response format: {json.dumps(result)}"
 
 def chat_with_models(user_input, history):
     responses = []